On Monday 8th of October, team DriverLeics packed their two Audi Q2 scale 1:8 cars and went for a great showdown to Ingolstadt, Germany.
Team DriverLeics had worked towards this competition for three and a half months in different buildings of the University of Leicester pursuing various approaches to autonomous driving. The qualification event allowed all participating teams to practice and fine-tune their software on a large track in Ingolstadt.
All teams had set up their workspaces next to the track for three days. Audi supplied food, drinks, electricity, and internet for everyone to work. It was a great atmosphere working among fifteen European teams of autonomous vehicle enthusiasts.
The students also had a chance for "speed dating" experts of the Audi group to discuss the latest technologies and opportunities for working with this Vorsprung durch Technik company.
Team DriverLeics had worked on different approaches for bringing the 1:8 scale Q2 onto the road. One of the innovations tried out by the team was accurate localization by comparing expected road markings with the actually observed ones. Another highlight of the software was the use of hierarchical state machines to describe different driving modes of the car, e.g., overtaking another car or stopping for a pedestrian. Of the teams we talked to, we were the only team applying model-based development with code generation for driving behaviors.
On the day of the competition the tension was high for all teams. In an order determined by draw the cars were placed on the track. Only the team leader was allowed to stand by the car with the jury while all other participants had to take spectator positions. Our car was the last one to compete on that day. After an emotional finish it was clear that team DriverLeics would not qualify for the finals. The first year with previously unknown technologies seems to be the hardest for all teams.
We hope to come back with such a great team in 2019!
Team DriverLeics had worked towards this competition for three and a half months in different buildings of the University of Leicester pursuing various approaches to autonomous driving. The qualification event allowed all participating teams to practice and fine-tune their software on a large track in Ingolstadt.
All teams had set up their workspaces next to the track for three days. Audi supplied food, drinks, electricity, and internet for everyone to work. It was a great atmosphere working among fifteen European teams of autonomous vehicle enthusiasts.
The students also had a chance for "speed dating" experts of the Audi group to discuss the latest technologies and opportunities for working with this Vorsprung durch Technik company.
Team DriverLeics had worked on different approaches for bringing the 1:8 scale Q2 onto the road. One of the innovations tried out by the team was accurate localization by comparing expected road markings with the actually observed ones. Another highlight of the software was the use of hierarchical state machines to describe different driving modes of the car, e.g., overtaking another car or stopping for a pedestrian. Of the teams we talked to, we were the only team applying model-based development with code generation for driving behaviors.
Our car was equipped with an artificial intelligence to recognize and identify pedestrians on the roadside. This AI had been trained using state-of-the-art deep learning technologies with countless images recorded from many driving scenarios. Finally, and to piece it all together the car was using a route planner, similar to one in a car's GPS, and a path planner for detailed planning of driving maneuvers.
On the day of the competition the tension was high for all teams. In an order determined by draw the cars were placed on the track. Only the team leader was allowed to stand by the car with the jury while all other participants had to take spectator positions. Our car was the last one to compete on that day. After an emotional finish it was clear that team DriverLeics would not qualify for the finals. The first year with previously unknown technologies seems to be the hardest for all teams.
We hope to come back with such a great team in 2019!
Comments
Post a Comment